Crime, Delinquency & Arrest

Crime, Delinquency & Arrest

Created Variables

FP_YYCRIMI. Provides Delinquency Index. To create this index, Child Trends, Inc. used questions in rounds 1 and 2 that asked respondents if they ever participated in various criminal/delinquent activities.  In rounds 2-4, Child Trends, Inc. used questions that asked respondents if they participated in various criminal/delinquent activities since the date of their last interview (FP_YYCRIMI_DLI) as part of the index. See Codebook Supplement Appendix 9 for more information and statistical analysis. 


Event History Variables

The arrest and incarceration event history arrays consist of monthly variables that document the number of arrests and incarcerations in each month starting at the respondent’s 12th birthday. Using these arrays, researchers can extract the status of a respondent at a point in time or over time. These arrays are also constructed to be similar to the NLSY97 event history arrays in other domains such as employment, schooling, marriage, cohabitation and fertility.


This array lists the respondent’s number of arrests on a monthly basis. It starts in January 1992 (the month in which the oldest respondent turned age 12) and ends with the most recent publicly available interview date. Arrests and incarcerations that occurred before the respondent turned 12 will not be included in this array.  The codes and their definitions are as follows:

  Code Definition
  -4 Assigned if R is younger than 12 years old or has not been interviewed about this month
  0 Assigned if R was not arrested in this month and previously was not arrested
  1-98 Indicates the number of times R was arrested in this month
  99 Assigned if R had been arrested previously but was not arrested in this month

Missing and imputed values.  Occasionally, respondents cannot provide information about arrest dates and the number of arrests. In early rounds of the survey, if respondents cannot provide the arrest date (both month and year) or the year of the arrest, the arrest is not populated in the arrest event history array. However, dates have been imputed for skipped arrests from round 7 onwards. In the main questionnaire, for respondents who reported 4 or more arrests since date of last interview, only the first and last arrest were dated. For the arrest arrays, the middle arrest dates were imputed as being evenly spaced between the first and last arrest dates.  Where first or last arrest year or month was missing it was imputed on a case by case basis based on last interview date and known arrest date information. 

If respondents cannot provide only the arrest month, then it is imputed using the month of the middle of the period since the last interview date. For example, if a respondent was interviewed in Round 6 and in Round 8, but not in Round 7, the program will take the month from the mid-date between Round 6’s interview date and Round 8’s interview date.  If respondents cannot provide arrest month in Round 1, the missing month is set to June.

Arrest summary file: An arrest summary data set was also created to provide summary measures of the respondent’s arrest event history.  For each respondent, this data includes the first arrest date reported, the total number of arrests (both dated and undated) and the number of arrests with missing year, month or both month and year.

  Variable Definition
  ARREST_FIRST Earliest arrest date as reported by R. If R did not provide an arrest date but was arrested, this is set to “-3”.
  ARREST_TOTNUM Total number of arrests as reported by R
  ARREST_MISSNUM Total number of rounds (question years) that R refused to answer the question on number of arrests since the date of last interview
  ARREST_DATED Total number of arrests with arrests dates (including missing months). This should equal the number of arrests in ARREST_STATUS array.
  ARREST_MISSINGDT Number of arrest dates with missing month and missing year
  ARREST_MISSINGYR Number of arrest dates with missing year
  ARREST_MISSINGMON Number of arrest dates with missing month
  ARREST_UNASKED Number of arrests where arrest date was not asked
  ARREST_LASTINTDATE Date of last interview with


This array lists the respondent's incarceration status on a monthly basis. It starts in January 1992 and ends with the most recent publicly available interview date. Once again, note that incarcerations that occurred before the respondent turned 12 will not be included in this array. Incarceration in the incarceration event history refers to jail or adult correctional facilities. Juvenile detention centers are not included. The codes and their definitions are as follows:

  -4 Assigned if R is younger than 12 years old in this month or has not been interviewed about this month
  0 Assigned if R was not incarcerated this month and previously never incarcerated
  1 Indicates R was incarcerated during all or some portion of this month
  99 Indication R was not incarcerated this month but has previously been incarcerated

Missing and Imputed Values. Occasionally, respondents cannot provide information about entry and exit months/years for incarcerations. Where these were missing, they were imputed using known/given prior and future arrest date information as well as prior and future interview dates.  For example, if a respondent indicated being actively incarcerated in one rounds interview and then not being incarcerated in the next interview, with no given exit date, the exit date was given as the mid-month between the interviews.  If, instead, the Respondent had a listed re-arrest date that was earlier than the next interview date, the mid-month between re-arrest and when the Respondent was last known to be incarcerated (the prior interview date) was used. 

Incarceration dates are only given in month/year. Hence, a one day and a 30-day incarceration could both be marked as extending from the same month to the same month, if the 30-day one started on the 1st and the respondent was released before the next month. For the summary values any incarceration starting and ending on the same month is counted as a one-month incarceration. All longer incarcerations are equally inclusive, counting both the entry month and exit month as full months. These could be almost an entire month shorter depending on exact entry/exit dates.

Incarceration summary file: An incarceration summary file provides summary measures of the respondent's incarceration history. The variables in this data file for each respondent include the date of last interview with the respondent, the first entry date into incarceration, the total number of separate incarceration spells, the age at first incarceration, the length of the first incarceration and longest incarceration as well as whether the respondent was currently incarcerated at the date of the last interview.

  Variable Definition
  INCARC_FIRST Earliest entry date into incarceration as reported by the R 
  INCARC_TOTNUM Total number of separate incarcerations reported by R
  INCARC_AGE_FIRST Age of R when first incarcerated
  INCARC _LENGTH_FIRST Months R was incarcerated the first time incarcerated
  INCARC_LENGTH_LONGEST Months R was incarcerated during R’s longest incarceration
  INCARC_TOTMONTHS Total months R has spent incarcerated
  INCARC_CURRENT Yes if R currently incarcerated at date of last interview

The NLSY97 survey questions on crime, delinquency, and arrest can be divided into two areas: questions about active participation and questions about perceptions of personal safety and other antecedents to delinquent behavior.  This section focuses on the youth's participation in delinquent and criminal behaviors; other details can be found in Household & Neighborhood Environment. The following information on delinquency, crime, and arrest is collected in the self-administered section of the youth instrument.

Within the criminal justice system, "delinquent" behavior is usually differentiated from "criminal" behavior on the basis of an offender's age.  For example, a 12-year-old who has damaged property would likely be charged with a delinquent offense; an older youth might be charged with either a delinquent or criminal offense at the discretion of the judge.  Since an offense is only classified as criminal or delinquent if the offender actually appears in court, a different distinction is made for the purposes of this discussion.  The descriptions of delinquency in this section include behaviors, such as running away from home, which are unlawful for minors but not necessarily for those ages 18 and older.  The discussions of crime and arrest in this section describe the questions on behaviors that are considered criminal activity for adults, even though a respondent who has committed one of these offenses may have actually been charged as a juvenile rather than an adult.


The first set of delinquency questions addresses the respondent's history of running away from home and staying away at least overnight. Data include the age when this first occurred, the total number of times the respondent has ever run away, and the total number of times he or she has run away since the date of last interview. After round 1, youths who were age 17 or older at the date of last interview or who were not living with any parents were skipped past the runaway questions. Questions about running away from home were discontinued after round 6.

In rounds 1 through 9, respondents answer a series of questions on whether they have ever carried a handgun, defined as any firearm other than a rifle or a shotgun, and if they have carried one since the last interview (starting in round 2). The series continues by asking their age the first time they did so, the number of days out of the past 30 that they carried the gun, and whether they carried the gun to school during this period.

In a separate set of questions, respondents are surveyed on their involvement in gang activity. Those who belonged to a gang are asked to state the age when they joined the gang and their membership status in the previous 12 months (round 1) or since the date of the last interview (rounds 2-9). Those who did not belong to a gang during that time period state their age when they last belonged to a gang. Respondent gang activity questions were last asked in round 9, although respondents have been asked through the current round about the presence of gangs in their neighborhood or school. Additional information on gang activity in the respondent's extended family, school, and neighborhood is found in Household & Neighborhood Environment.


The survey asks about participation in and the intensity of various criminal activities in the previous year (round 1) or since the date of last interview (subsequent rounds). The types of criminal activities included are:

  • Purposely damaged or destroyed property not belonging to the respondent
  • Stole something worth less than $50
  • Stole something worth $50 or more (including a car)
  • Other property crimes, including fencing stolen property, possessing or receiving stolen property, or selling something for more than it was worth
  • Attacked or assaulted someone
  • Sold or helped to sell marijuana, hashish, or other hard drugs

Youths who stole something state whether they stole from a store, snatched someone's purse/wallet or picked someone's pocket, went into a locked building to steal something, or used a weapon during the act. Additionally, respondents who report stealing something worth $50 or more are asked if they stole a car.

In round 1, follow-up questions about a particular activity (except stealing something worth less than $50) determined the respondent's age the first time and the number of times he or she participated in the past 12 months. Respondents who reported no involvement during the previous 12 months were asked to state their age when they last participated in that criminal activity. If the respondent committed other property crimes, sold or helped to sell drugs, or stole something worth $50 or more, he or she was asked about any monetary rewards in 1996--the total cash received or the total cash he or she would have received--from selling these items. Respondents who reported having sold or helped to sell drugs stated whether they were involved in selling marijuana or hashish, other hard drugs, or both, as well as the amount of income received from these sales. 

Reference periods and question details shifted in round 2. Subsequent rounds followed the same pattern as round 2. In these surveys, respondents were questioned about the number of times they participated in each activity since the date of last interview. However, respondents who answered "don't know" or "refused" in round 1 were again asked about criminal activities in which they had ever participated.  Those who stole something (worth more or less than $50) or participated in the other property crimes listed above were asked for the amount of any monetary rewards in the previous calendar year--the total cash received or the total cash he or she would have received--from selling these items. Respondents who reported selling or helping to sell drugs distinguished whether they were involved in selling marijuana or hashish, other hard drugs, or both, and listed the amount of income received from these activities. Subsequent rounds followed the same pattern as round 2. 

Beginning in round 8, the universe for the crime section was restricted. The crime questions were asked only of respondents who had ever reported being arrested and also a control group of approximately 10% of the respondents for comparison.

Arrests, Incarceration, and Post-Incarceration

NLSY97 youth respondents are asked whether they have ever been arrested by the police or taken into custody for an illegal or delinquent offense (not including arrests for minor traffic violations) and the total number of times this has happened. Round 1 then collected information on the number of arrests before the respondent's 12th birthday and the respondent's age at the first arrest. Further data were gathered on the respondent's most recent and first arrests.  (If the respondent reported only one arrest, this information was collected as the respondent's most recent arrest; the first arrest data fields contain missing values.) 

Rounds 2 through the current round gathered the number of arrests since the last interview, including the month and year of the first arrest if it occurred during that time. Note: for rounds 2-4, the design of the arrest loop collected information only on arrests since the date of last interview. If a respondent had just been arrested but had not yet been charged/convicted/sentenced prior to the interview, there was no effort during these rounds to collect this information in the next interview. Beginning in round 5, questions were asked about pending outcomes of arrests from the prior round in order to determine their final outcome regarding any charges, convictions, and sentences.

For the first and most recent arrests (round 1) or a continuous loop of all arrests (subsequent rounds), the respondent provides the month and year of the arrest and states whether the police charged him or her with an offense. A series of questions then determines the type of offenses with which the respondent was charged; he or she could enumerate multiple charges. The list of possible arrest charges includes assault, burglary, destruction of property, possession or use of illicit drugs, sale or trafficking of illicit drugs, a major traffic offense, and a public order offense. 

For each offense with which they have been charged, respondents report whether they were convicted for or pled guilty to that offense. A follow-up question distinguishes whether the respondent went to juvenile court, adult court, both, or neither as a result of the charges.  Respondents then describe the sentence. 

If the respondent did not go to court, was not convicted of, and did not plead guilty to any charges, he or she states the arrest outcome: no further action, pre-court diversion program/counseling, appearance before a judge with no further action, or other.

For those sentenced to jail, an adult corrections institution, a juvenile corrections institution, or a reform/training school, the surveys record the month and year they began their sentence and the month and year their sentence ended (if appropriate). Similar questions ask about the month and year community service or probation began and ended. Questions were also asked in 2002 and in 2007-2009 about any member of R's household being in jail or prison.

Starting in round 12, respondents answered a more detailed series of questions about incarceration, including both current and past incarceration. They were asked about the number of visits, phone calls, and amount of mail they received while incarcerated and how far the institution was from their home. Respondents also provided information about losing a job or child custody because of incarceration. Several questions were asked about any rehabilitation services respondents received while incarcerated, such as re-entry preparations, substance abuse treatment, mental health treatment, medical attention, employment training, GED or other schooling, and religious services or ministry.

Respondents also provided information (starting in round 12) about post-incarceration activity, including where the respondent spent his/her first night following release, any problems finding a place to live, how hard or easy it has been to stay out of prison or not to commit crimes, and the respondent's estimate of the likelihood of being re-incarcerated within the next five years. In addition, information was gathered on parole restrictions, including where the parolee could live and travel, hours away from home, and with whom the parolee could associate.

Design Effects for Ever-Arrested Variables

As discussed in the Sample Weights and Design Effects section, the initial clustering of the sample can increase standard errors relative to a simple random sample. Table 1 gives design effects for the "HAS R BEEN ARRESTED?" variable from three select years. The table shows that the design effect is largest for Round 1 (when the clustering of the sample was greatest), but substantially smaller in later years.

Table 1. Design Effects for Arrest Variable, Select Years

Round Variable Sample Size Mean SE Mean DEFF Mean
 7 HAS R BEEN ARRESTED SDLI? 2003 7729  0.05  0.0028  1.22
 13 HAS R BEEN ARRESTED SDLI? 2009 7509  0.03  0.0022  1.14

Other Topics Related to Crime

A description of the questions on the attitudes of the respondent toward justice is found in Attitudes; the use of drugs by the respondent is described in Drug Use; and descriptions of survey questions on the safety of the respondent's household environment are found in Household & Neighborhood Environment.

Comparison to Other NLS Surveys: Information on crime, delinquency, and arrest records was collected from the NLSY79 in a special self-report supplement during the 1980 interview. This supplement detailed respondents' participation in and income from criminal activities and their contacts with the criminal justice system.  Starting in 1988, the surveys of the Children of the NLSY79 asked those age 10 and older (including the Young Adults) to report on participation in various illegal activities.  For more information, consult the NLSY79 Child & Young Adult User's Guide.

Survey Instruments: Questions on crime, delinquency, and arrest are found in the self-administered section (question names begin with YSAQ) of the Youth Questionnaire.

Related User's Guide Sections Attitudes
Drug Use
Household & Neighborhood Environment
Main Area of Interest Illegal Activity & Arrest
Family Process Measures
Supplemental Areas of Interest Expectations
Household Characteristics
Substance Use